Butane oxidation process development in a circulating fluidized bed

https://doi.org/10.1016/j.apcata.2009.10.023Get rights and content

Abstract

DuPont designed and operated a circulating fluidized bed reactor (CFB) to produce maleic anhydride from n-butane using a vanadium pyrophosphate catalyst (VPP) encapsulated in a silica shell. A fraction of the pyrophosphate was oxidized to the V5+ state from the V4+ state in an air fed fluidized bed regenerator. The oxidized VPP was shuttled to a transport bed reactor with a high concentration of butane and oxygen. The gas carried the catalyst up through the bed at velocities of 0.8 m/s and, in the commercial plant, solids circulation rates exceeding 7 kt/h. Early development work was conducted on an experimental scale facility containing 1 kg of catalyst. The pilot plant catalyst inventory exceeded 2000 kg and there was 175 t in the commercial reactor. Throughout the program, significant advances in catalyst manufacture and process design were achieved. The CFB reactor configuration is being considered for several unrelated processes including chemical looping combustion, methanol-to-olefins and hot gas desulphurization. Improvements in spray drying technology reduced attrition losses by an order of magnitude versus expectation based on the pilot plant. Together with the low attrition losses and good stability, catalyst consumption was reduced by successfully re-spray drying used/attrited catalyst. By modifying the solids entrance and exit configurations, we were able to double initial plant capacity. Operability of the plant was excellent with a turn-down ratio of 5 demonstrated. At a production rate of 65,000 t/year of maleic acid – one of the largest single train reactors for a partial oxidation of an alkane – the maximum temperature difference within the bed was less than 20 °C. Heat transfer had been a major design consideration but even at this rate, only 1/3 of the total coil surface for cooling was activated.

Graphical abstract

DuPont designed and operated a circulating fluidized bed reactor to produce maleic anhydride from n-butane using a vanadium pyrophosphate catalyst encapsulated in a silica shell. Challenges overcome during the commercialization phase included: attrition resistance, heat transfer, and oxygen addition to a vessel with a high butane concentration. Maintaining the catalyst oxidized was critical to catalyst activity and maleic anhydride selectivity.

  1. Download : Download high-res image (88KB)
  2. Download : Download full-size image

Introduction

The most common route to produce THF – an industrial solvent as well as a critical monomer for segmented polyurethanes (Lycra®) and copolyester elastomers (Hytrel®) – has been the Reppe process: acetylene reacts with formaldehyde forming butynediol followed by hydrogenation and an acid de-hydration step. DuPont developed a new environmental process to THF in which n-butane is partially oxidized to maleic anhydride and the acid is subsequently hydrogenated at high pressure [1], [2]. The first step – partial oxidation of butane over a vanadium phosphorous oxide (VPO) – was accomplished through major innovations in reactor technology and catalyst morphology.

The circulating fluidized bed reactor (CFB) transferred catalyst from a net oxidizing zone to a net reducing zone. In the oxidizing environment, a fraction of the vanadyl pyrophosphate (VPP) – (VO) 2P2O7 – was converted from a V4+ oxidation state to V5+. In the reducing environment, the V5+ phase was a source of oxygen to partially oxidize n-butane to maleic anhydride. In multi-tubular fixed beds the reactants are co-fed and operate under very oxidizing conditions below the lower explosion limit. Circulating catalyst from an oxidizing environment to a reducing environment has been shown to improve selectivity to maleic anhydride as well as activity—increased butane conversion [3], [4], [5]. Even in fixed bed operation, when the butane feed is temporarily interrupted and fed air, performance is improved when the butane is initiated. Schuurman and Gleaves [3] proposed that this may be evidence that more than one VPP phase may be active. Recently, Ballarini et al. [6] inferred the nature of the surface active phases by running transient experiments. They showed that both the P/V ratio and reaction conditions could alter the surface to form vanadium oxide and polyphosphoric acids or other non-selective active species.

Together with higher selectivity and increased yield, the net throughput per kg of catalyst is improved in a CFB by operating with more than 10  vol% butane, which is 2–5 times as much as a turbulent fluidized bed and multi-tubular fixed bed, respectively. Hutchings [7] suggested that by increasing the butane concentration, the maleic anhydride yields could be increased substantially. Centi et al. [8] reported higher selectivity to MA (as well as butenes, butadiene and other by-products) with 30% butane concentrations in the feed compared to 1% at low temperatures (300 °C) and 10% oxygen. However, as part of the development of a membrane reactor process, Mota et al. [9] fed high concentrations of butane to a fixed bed reactor to simulate the entrance conditions and found a rapid reduction of the catalyst surface. The V5+ species were reduced to V4+ together with a significant carbon build up and a corresponding reduction in activity. Reoxidizing the catalyst for an extensive period of time (2 h) reconstituted the surface. Mallada et al. [10] fed a VPP catalyst 10% butane and found that both selectivity and yield rose while increasing the inlet oxygen concentration from 10% to 26%. Under high n-butane inlet feed concentrations, Ballarini et al. [11] reported that maleic anhydride selectivity is lower when the oxygen conversion is near complete due to higher carbon oxides and tetrahydrophthalic and phthalic anhydrides. However, when the reactor was run under partial oxygen conversion, selectivity was improved. Lorences et al. [12] mapped out the relationship between by-product acids – acetic and acrylic acids predominated but others include fumaric, methacrylic, and phthalic acids—butane and oxygen feed concentrations. They showed that as the oxygen conversion approached 100%, the maleic anhydride selectivity dropped as did the overall butane reaction rate while the by-product acid selectivity increased.

High concentrations of butane result in increased maleic anhydride yield as long as the catalyst does not become over reduced and transferring to an oxidizing environment can maintain the overall activity. Shuttling catalyst between two zones – oxidation and reduction – leads to high mechanical stresses on the catalyst micro-spheres. Commercializing this process requires catalyst with a high degree of attrition resistance. Bergna [13] developed spray drying technology to impart attrition resistance while at the same time minimizing the binder concentration. Often, the binder constitutes up to 50% or more of the total mass. Precursor was produced using an organic solvent route with iso-butanol and benzyl alcohol. After washing and drying, the 60 μm precursor powder was milled to between 1 and 2 μm. The milled powder was slurried together with polysicilic acid and spray dried resulting in porous microspheres with an average diameter of 70 μm. The microspheres were calcined and activated with air at pressures up to 6 bar and 390 °C for 4 h [14]. This procedure was followed for the initial charge of catalyst but further developmental work allowed us to calcine and activate in situ under operating conditions. Albonetti et al. [15] describe a similar thermal treatment method to activate catalyst but required 1000 h of on-line operation with butane to achieve a steady “equilibrated” catalyst.

The calcination and activation procedure was an example of successfully developing an experimental protocol and implementing it in the commercial operation. In this work, we describe the reactor performance data collected in the commercialization phase of DuPont's CFB process to produce maleic anhydride/acid: the reactor geometry at three scales – experimental, pilot and commercial – are described in detail together with experimental data that characterizes their performance under standard operating conditions. Based on these data, we show the relationship between maleic anhydride/acid selectivity as a function of butane and oxygen concentration as well as the lattice oxygen contribution.

Section snippets

1/4″ riser catalyst qualification protocol

Experiments to qualify catalyst and test stability were conducted in a 1/4″ riser facility that consisted of 4 interconnected quartz vessels housed in an electric furnace, as shown in Fig. 1: fast bed, riser, stripper and regenerator. A total of 1 kg of VPP was charged to the reactor with approximately 600 g in the regenerator, 300 g in the stripper and the 60–80 in the fast bed/riser. In the standard protocol to evaluate catalyst performance, the reactor was operated at three temperatures – 360,

Laboratory evaluations

The major focus of the experimental facilities after the initial exploratory phase was to qualify precursor for the pilot plant and commercial reactors, toxicity testing, process modelling and developing protocols for calcination and activation under industrial conditions [16]. Qualifying precursor and catalyst involved measuring performance parameters such as activity, selectivity, lattice oxygen capacity, and attrition resistance as well as certain physico-chemical properties including

Conclusions

Commercializing butane oxidation technology in a CFB was a challenging endeavor that entailed advances in both catalyst manufacture and process technology. Developing an attrition resistant catalyst with only as little as 10% silica binder resulted in smaller reactor volumes and significant economies with respect to catalyst consumption. Attrition rates in the plant dropped to as low as 1 kg/h whereas, based on pilot plant experience, the expectation was attrition rates as high as 15 kg/h. The

References (19)

  • R.M. Contractor

    Chem. Eng. Sci.

    (1999)
  • Y. Schuurman et al.

    Catal. Today

    (1997)
  • N. Ballarini et al.

    Catal. Today

    (2006)
  • G.J. Hutchings

    Appl. Catal.

    (1991)
  • G. Centi et al.

    J. Catal.

    (1984)
  • S. Mota et al.

    J. Catal.

    (2000)
  • R. Mallada et al.

    Catal. Today

    (2000)
  • N. Ballarini et al.

    Catal. Today

    (2005)
  • S. Albonetti et al.

    J. Catal.

    (1996)
There are more references available in the full text version of this article.

Cited by (0)

View full text